Ai-is-light
commented
6 years ago Creating layer global_pool I0823 11:47:52.993285 22 net.cpp:99] Creating Layer global_pool I0823 11:47:52.993289 22 net.cpp:479] global_pool <- res5b_bn I0823 11:47:52.993295 22 net.cpp:130] This layer is inheriting previous layer's sync mode: 1 I0823 11:47:52.993300 22 net.cpp:435] global_pool -> global_pool I0823 11:47:52.993306 22 net.cpp:163] Setting up global_pool F0823 11:47:52.993738 22 blob.cpp:32] Check failed: shape[i] >= 0 (-1 vs. 0)
how set the following layer: ######################################### Global pool ############################################################ layer { name: "global_pool" bottom: "res5b_bn" top: "global_pool" type: "Pooling" pooling_param { pool: AVE kernel_size: [4, 7, 7] stride: [1, 1, 1] } }
mzolfaghari
@mzolfaghari Hi, I plan to train the custom dataset with the different "num_segments" for the speed & accuracy. And, I change the following information: (1) layer { name: "data" type: "VideoData" top: "data" top: "label" video_data_param {
source: ".../train.txt"
} transform_param{ crop_size: 224 mirror: true fix_crop: true more_fix_crop: true multi_scale: true max_distort: 1 scale_ratios:[1,.875,.75, .66] is_flow: false mean_value: [104] mean_value: [117] mean_value: [123]
} include: { phase: TRAIN } }
(2)
=================== 3D network ===========================================
layer {
name: "r2Dto3D"
type: "Reshape"
bottom: "inception_3c_double_3x3_1_bn"
top: "res2b_bn_pre"
reshape_param{
shape { dim: -1 dim: 16 dim: 96 dim: 28 dim: 28 }
}
}
layer { name: "r2Dto3D" type: "Reshape" bottom: "inception_3c_double_3x3_1_bn" top: "res2b_bn_pre" reshape_param{ shape { dim: -1 dim: 8 dim: 96 dim: 28 dim: 28 } } }
(3)
layer { name: "reshape_fc_st2" type: "Reshape" bottom: "global_pool2D" top: "reshape_fc_st2" #reshape_param { shape { dim: [-1, 1, 16, 1024] } } }
layer { name: "segment_consensus_st2" type: "Pooling" bottom: "reshape_fc_st2" top: "pool_fusion_st2" #pooling_param { pool: AVE kernel_h: 16 kernel_w: 1 } }
layer { name: "reshape_fc_st2" type: "Reshape" bottom: "global_pool2D" top: "reshape_fc_st2" reshape_param { shape { dim: [-1, 1, 8, 1024] } } } layer { name: "segment_consensus_st2" type: "Pooling" bottom: "reshape_fc_st2" top: "pool_fusion_st2" pooling_param { pool: AVE kernel_h: 8 kernel_w: 1 } }
However, I got the following error; I0823 11:47:52.993273 22 net.cpp:170] Top shape: 1 1024 (1024) I0823 11:47:52.993278 22 layer_factory.hpp:74] Creating layer global_pool I0823 11:47:52.993285 22 net.cpp:99] Creating Layer global_pool I0823 11:47:52.993289 22 net.cpp:479] global_pool <- res5b_bn I0823 11:47:52.993295 22 net.cpp:130] This layer is inheriting previous layer's sync mode: 1 I0823 11:47:52.993300 22 net.cpp:435] global_pool -> global_pool I0823 11:47:52.993306 22 net.cpp:163] Setting up global_pool F0823 11:47:52.993738 22 blob.cpp:32] Check failed: shape[i] >= 0 (-1 vs. 0) Check failure stack trace: @ 0x7f3e01c5c5cd google::LogMessage::Fail() @ 0x7f3e01c5e433 google::LogMessage::SendToLog() @ 0x7f3e01c5c15b google::LogMessage::Flush() @ 0x7f3e01c5ee1e google::LogMessageFatal::~LogMessageFatal() @ 0x7f3e01ff7878 caffe::Blob<>::Reshape() @ 0x7f3e020f5a7b caffe::PoolingLayer<>::Reshape() @ 0x7f3e0206deec caffe::CuDNNPoolingLayer<>::Reshape() @ 0x7f3e02135abd caffe::Net<>::Init() @ 0x7f3e0213821c caffe::Net<>::Net() @ 0x4118a3 time() @ 0x40e932 main @ 0x7f3e0066f830 __libc_start_main @ 0x40ef79 _start @ (nil) (unknown) Aborted (core dumped)
Looking forward to any replies. Thanks